A very young, compact bipolar H₂O maser outflow in the intermediate-mass star-forming LkHα 234 region

Abstract

1 INTRODUCTION

The early stages of evolution of low-mass stars are relatively well characterized by the formation of a system with a central protostar, accreting material from a rotating accretion disc at scales of a few hundreds of astronomical units (au), and simultaneously ejecting a collimated outflow with the presence and crucial role of magnetic fields. Accretion and mass-loss processes, two closely related mechanisms, govern the formation of low-mass stars (e.g. Girart, Rao & Marrone 2006; McKee & Ostriker 2007; Machida, Inutsuka & Matsumoto 2008; Armitage 2011; Williams & Cieza 2011). In fact, spatio-kinematical studies of jets and Herbig–Haro (HH) objects show that outflows from low-mass young stellar objects (YSOs) are non-steady, but present variability in the ejection velocity as well as pulsed events, probably related to recurrent instabilities in the accretion discs (e.g. Zinnecker, McCaughrean & Rayner 1998; Reipurth & Bally 2001; Estalella et al. 2012). With respect to the formation of massive stars (≳10 M_⊙), there are several examples showing the presence of massive disc–protostar–jet systems at scales of thousand au (in some cases with magnetic fields oriented parallel to the collimated outflows), indicating that stars at least up to ∼20 M_⊙ form via an accretion disc as low mass do (e.g. Patel et al. 2005; Jiménez-Serra et al. 2007; Torrelles et al. 2007; Carrasco-González et al. 2010, 2012b; Davies et al. 2010; Vlemmings et al. 2010; Fernández-López et al. 2011; Maud & Hoare 2013). Furthermore, variation in the ejection velocity of outflows as well as remarkable short-lived episodic ejection events (few tens of years) have been reported towards massive YSOs through radio continuum (e.g. Martí, Rodríguez & Reipurth 1995; Curiel et al. 2006) and H₂O maser spatio-kinematical observations (e.g. Torrelles et al. 2001, Torrelles et al. 2003; Surcis et al. 2011a,b; Chibueze et al. 2012; Sanna et al. 2012; Kim et al. 2013; Trinidad et al. 2013). This indicates that outflows in massive YSOs are also non-steady, probably due to instabilities in the accretion processes as in the case of low-mass YSOs.

Mainly because most of the observational efforts have been concentrated in low- and high-mass YSOs, little is known on what happens in intermediate-mass protostars, with only a few examples showing clear disc–YSO–jet systems at scales of a few hundred au (e.g. Palau et al. 2011; Carrasco-González et al. 2012a), but up to now without any observational information on the magnetic field and outflow kinematical behaviour at these small scales. Observations of the first stages of the evolution of intermediate-mass YSOs at small scales are therefore important to fill that gap, and thus to have a more complete vision about the processes of star formation throughout a more continuous range of masses. In this paper, we present multi-epoch Very Long Baseline Interferometry (VLBI) H₂O maser observations towards intermediate-mass YSOs to study the spatio-kinematical distribution of their masers with ∼0.5 mas resolution. This kind of study has previously been very useful in massive star-forming regions to detect outflow activity through the presence of H₂O maser emission, allowing the identification of new centres of star formation, some of them are associated, as mentioned before, with short-lived episodic ejection events with kinematic ages of a few tens of years. With this work, we now extend this study to intermediate-mass star-forming regions.

The optically visible Herbig Be star LkHα 234, with luminosity ∼10³ L_⊙ and mass ∼5 M_⊙, is one of the brightest sources of the NGC 7129 star-forming region, located at a distance of 0.9 kpc (Herbig 1960; Strom et al. 1972; Bechis et al. 1978; Fuente et al. 2001; Trinidad et al. 2004; Xu et al. 2013). High angular resolution near- and mid-infrared images obtained by Kato et al. (2011) reveal, in addition to LkHα 234, a cluster of eight YSO candidates within a distance of ∼10 arcsec from LkHα 234 (named by Kato et al. as objects B, C, D, E, F, G, NW1, and NW2). Very Large Array (VLA) observations show five radio continuum sources in a region of ∼5 arcsec (VLA 1, VLA 2, VLA 3A, VLA 3B, LkHα 234; Trinidad et al. 2004), some of them with mid-infrared counterparts: VLA 2 (NW2), VLA 3A+3B (NW1), and LkHα 234 (see Fig. 1). NW2 and NW1 are not seen in the J-, H-, and K-band images, but they are bright in the L′, M′, and mid-infrared bands, indicating that they are highly embedded YSOs (Kato et al. 2011). H₂O maser emission is detected towards this region, mainly associated with VLA 1, VLA 2 (NW2), and VLA 3A+3B (NW1) (Tofani et al. 1995; Umemoto et al. 2002; Trinidad et al. 2004; Marvel 2005; see Fig. 1). Several outflows emerging from this region have been also observed, e.g.: (i) a jet observed in [S ii], extending to the south-west (p.a. ∼ 252°) up to distances of ∼40 arcsec (∼36 000 au), without evidence of a counter-jet (Ray et al. 1990); (ii) an infrared jet seen in v=1–0 S(1) H₂ emission, extending towards the south-west (p.a. ∼ 226°) up to distances of ∼10 arcsec (∼9000 au), also without any evidence of a counter-jet (Cabrit et al. 1997). The overlapping of these two outflows within a small area of the sky makes it very difficult to clearly distinguish between various candidate driving sources. What we know, however, is that all the different studies exclude that the Herbig Be star LkHα 234 is driving any of the outflows of this star-forming region, but that the main outflow activity centres are probably the YSOs VLA 2 (NW2) and VLA 3A+3B (NW1) (Trinidad et al. 2004; Marvel 2005; Kato et al. 2011). Following the nomenclature of Trinidad et al. (2004), in this paper, we will refer to the region containing all this cluster of YSOs as the ‘star-forming LkHα 234 region’.

Figure 1.

JHK colour composite image (top-left panel), and 11.8 μm and 17.6 μm Keck LWS images (top-right panels) around the Herbig Be star LkHα 234. In addition to LkHα 234, eight YSO candidates are also indicated (B, C, D, E, F, G, NW1, NW2; images and nomenclature from Kato et al. 2011). Bottom panel: VLA-3.6 cm continuum contour map of the region showing emission from four sources, VLA 1, VLA 2, VLA 3 (this source, when observed with higher angular resolution, has two components, VLA 3A and 3B, which are separated by ∼0.3 arcsec; Trinidad et al. 2004), and LkHα 234 (the optical position is indicated by a star). The crosses indicate the position of the H₂O masers detected with the VLA (epoch 1999 June 29; observations and map from Trinidad et al. 2004). The mid-infrared sources NW1 and NW2 detected by Kato et al. (2011) coincide in position with the radio continuum sources VLA 3 (A+B) and VLA 2, respectively.

Open in new tabDownload slide

In Section 2, we present the multi-epoch VLBI H₂O maser observations towards the star-forming LkHα 234 region and the main results, in particular, the discovery of a very young, compact bipolar H₂O maser outflow associated with VLA 2 (NW2). The implications of these results are discussed in Section 3, while the main conclusions of this work are presented in Section 4.

2 OBSERVATIONS AND RESULTS

The 6₁₆-5₂₃ H₂O maser transition (rest frequency = 22 235.08 MHz) was observed with the Very Long Baseline Array (VLBA) of the National Radio Astronomy Observatory (NRAO)¹ towards the star-forming LkHα 234 region at three epochs (2001 December 2, 2002 February 11, and 2002 March 5) during an observing on-source time of ∼5 h per epoch. A bandwidth of 8 MHz sampled over 512 spectral line channels of ∼0.21 km s⁻¹ width and centred at V_LSR = −9.3 km s⁻¹ was used. The data were correlated at the NRAO Pete V. Domenici Science Operations Center. Calibration and imaging was made using the Astronomical Image Processing System (aips) package of NRAO. Delay and phase calibration was provided by the sources 3C345, 3C454.3, BL Lac, B1739+522, and B2007+777, while bandpass corrections were determined through observations of 3C345, 3C454.3, and BL Lac. The size of the synthesized beam was ∼0.4 mas for the three epochs. Self-calibration of the uv data was made through a strong point-like maser persisting in all three epochs. This maser, with flux density ∼10 Jy and radial velocity V_LSR = −10.1 km s⁻¹, has absolute coordinates α(J2000.0) = 21^h43^m06|$.\!\!^{\rm s}$|312, δ(J2000.0) = 66°06′55|${^{\prime\prime}_{.}}$|79 (±0|${^{\prime\prime}_{.}}$|05), and is spatially associated with the VLA 2 source. This maser was also used for a first preliminary coordinate alignment of the three epochs of observations, which helped us to the identification of the H₂O maser emission in the region.

We detected H₂O maser emission towards the sources VLA 1, VLA 2, and VLA 3A+3B (Fig. 2). For each epoch, we obtained individual data cubes of 8096×8096 pixels × 512 velocity channels, and pixel size of 0.1 mas, for three fields, centred on each of the radio continuum sources. The rms noise level of the individual velocity channel maps depends on the intensity of the masers in each velocity channel, and ranges from ∼7 to 25 mJy beam⁻¹. For the identification of the maser spots in the region (by maser spot we mean emission occurring at a given velocity channel and distinct spatial position), we adopted a signal-to-noise ratio (SNR) threshold of ≳10. All the maser spots detected in the region, most of them unresolved with the beam size of ∼0.4 mas, were fitted with two-dimensional Gaussian profiles, obtaining their positions, flux density, and radial velocity. From the SNR and the beam size, we estimate that the 1σ accuracy in the relative positions of the maser spots within each epoch is ≲0.02 mas (beam/[2×SNR]; Meehan et al. 1998).

Figure 2.

H₂O maser spectra observed with the VLBA towards VLA 1 (top), VLA 2 (middle), and VLA 3A+3B (bottom) on 2002 February 11. These spectra were obtained by adding components from the clean images of the different regions with the task ISPEC (aips). For VLA 1, H₂O maser emission is only detected in this particular epoch. For VLA 3A+3B, weak emission (∼0.09 Jy) is also detected at V_LSR ≃ −37 km s⁻¹ in epoch 2001 December 2.

Open in new tabDownload slide

For the proper motion measurements of the H₂O masers, we have identified a set of maser features persisting in different epochs of our observations. By ‘maser feature’ we mean a group of spots with positions coinciding within a beam size as well as radial velocities appearing in consecutive velocity channels (typically within ∼1 km s⁻¹). We first measured the proper motion of the maser features adopting the maser spot used to self-calibrate the uv data as the reference for the preliminary alignment of the three epochs. Those first steps already showed a clear symmetric bipolar outflow distribution of masers around VLA 2 (see Section 2.1.2). However, since the initial reference maser spot may have its own proper motion, resulting in arbitrary offsets in the proper motions for the whole system, we have defined a more ‘stationary’ reference frame assuming that the mean position of the maser features around VLA 2 persisting in all three observed epochs remains unchanged with time. All the proper motion values given in this paper have been determined after this realignment by a linear fitting to the position of the maser features of the different epochs as a function of time (Table 1).

2.1 Results

2.1.1 Spatial distribution of the H₂O masers

H₂O maser emission was detected towards the sources VLA 1, VLA 2, and VLA 3A+3B in the three observed epochs, except in VLA 1, where emission was only detected on 2002 February 11. As previously reported by Trinidad et al. (2004) and Marvel (2005), there is no H₂O maser emission towards the Herbig Be star LkHα 234, consistent with the absence of outflow activity in this object. The maser emission spans a velocity range and flux density from V_LSR ≃ −13 to −12 km s⁻¹, S_ν ≃ 0.07 to 0.3 Jy (VLA 1), V_LSR ≃ −18 to −9 km s⁻¹, S_ν ≃ 0.07 to 11 Jy (VLA 2), and V_LSR ≃ −37 to −4 km s⁻¹, S_ν ≃ 0.07 to 13 Jy (VLA 3A+3B). In Fig. 2, we show the spectra of the H₂O maser emission obtained towards VLA 1, VLA 2, and VLA 3A+3B on 2012 February 11, while the positions and radial velocity distribution of the masers for the same epoch are plotted in Fig. 3 for a general overview of the region. In general, the spatial distribution of the masers that we find with the VLBA agrees with the spatial distribution found by Trinidad et al. (2004) with the VLA on 1999 June 29 but with a beam size of ∼ 0.08 arcsec (see Figs 1 and 3). Specially remarkable is the similarity of the structure found around VLA 2, both with the VLBA (Fig. 3) and the VLA (Fig. 1), with two main groups of masers separated by ∼0.4 arcsec (∼360 au) as well as maser emission at the centre of these two groups coinciding very well with the centre of the centimetre emission of VLA 2. We do not find a significant segregation in the radial velocity of the north-eastern and south-western groups of masers associated with VLA 2, with the two groups having similar radial velocities.

Figure 3.

Positions of the H₂O masers measured with the VLBA for epoch 2002 Feb 11. Colour code indicates the radial velocity of the masers in km s⁻¹. The positions of the peak emission of the radio continuum sources VLA 1, VLA 2, VLA 3A, VLA 3B at 3.6 cm (Trinidad et al. 2004) are indicated by open squares. No water maser emission is detected towards the star LkHα 234, located at ∼ 2 arcsec south-east from VLA 3A (outside of the region represented in this figure). The (0,0) has absolute coordinates α(J2000) = 21^h43^m06|$.\!\!^{\rm s}$|323, δ(J2000) = 66°06′55|${^{\prime\prime}_{.}}$|92 (± 0|${^{\prime\prime}_{.}}$|05). The accuracy in the relative positions of the maser spots is better than ∼0.02 mas, while the accuracy in the relative positions of the masers and the radio continuum sources plotted in this figure is ∼10–20 mas (see Section 2).

Open in new tabDownload slide

In this paper, all the offset positions of the H₂O masers are relative to the maser spot position (0,0) detected with the VLBA at the centre of the two groups of masers around VLA 2 in the first epoch of our observations (2001 December 2), with V_LSR ∼ −10 km s⁻¹, flux density ∼0.2 Jy, and absolute coordinates RA(J2000) = 21^h43^m06|$.\!\!^{\rm s}$|323, Dec.(J2000)= 66°06′55|${^{\prime\prime}_{.}}$|92 (±0|${^{\prime\prime}_{.}}$|05). We identify this particular maser detected with the VLBA, with the VLA maser reported by Trinidad et al. (2004) at the centre of these two groups of masers (∼1 Jy beam⁻¹), with similar velocity (∼−11 km s⁻¹) and absolute coordinates (RA[J2000] = 21^h43^m06|$.\!\!^{\rm s}$|322, Dec.[J2000] = 66°06′55|${^{\prime\prime}_{.}}$|95), and which is located very close to the peak position of the VLA 2 source (within ∼10 mas).² This also allows us to plot in Fig. 3 the relative positions of the different radio continuum sources (VLA 1, VLA 2, VLA 3A, VLA 3B) with respect to the VLBA masers with an accuracy of 10–20 mas, which is high enough for a proper comparison of the maser distribution with respect to the radio continuum sources. In this way, we see from Fig. 3 that, in addition to VLA 1 and VLA 2, both components of VLA 3 (3A and 3B) have associated H₂O maser emission. There is also maser emission in between both binary components, but with the present data we cannot establish whether it is associated with either component or if it is associated with a different YSO not yet detected.

2.1.2 Proper motions of the H₂O masers

We measured proper motions of 41 H₂O maser features in the region, of which 26 were detected in the three epochs, while 15 were detected in only two epochs. In Table 1, we list the main parameters of these features, including the detected epochs, radial velocity, intensity, offset positions, proper motion values with their uncertainties, and their association with any of the sources in the region. Most of the maser features where we measured proper motions are associated with VLA 2 (32), while nine are associated with VLA 3A+B. Proper motion measurements of the masers towards VLA 1 were not possible since they were detected in only one epoch (2002 February 11).

The spatial distribution and proper motions of the masers around VLA 2 show a very compact symmetric bipolar outflow in the north-east–south-west direction, consisting of two bow-shock-like structures separated by ∼0.4 arcsec (∼360 au) moving in opposite directions, with VLA 2 located between them. In Fig. 4, we show the spatio-kinematical distribution of the masers around VLA 2, with the radial velocities indicated in a colour scale (V_LSR ≃ −18 to −9 km s⁻¹), and the proper motion vectors represented by arrows. The magnitude of the proper motions is in the range of ∼5–30 km s⁻¹, although most of them have values ≳20 km s⁻¹ (see Table 1 and Fig. 4). Font, Mitchell & Sandell (2001) report an ambient molecular cloud velocity V_LSR ≃ −10.3 km s⁻¹, obtained from ¹³CO and C¹⁸O with a beam size of ∼14 arcsec. This velocity is similar to the mean radial velocity of the masers associated with VLA 2 (V_LSR ≃ −11.6 km s⁻¹; Table 1). Considering this, and that our data do not show any significant difference between the radial velocity of the north-eastern and south-western structures (although Trinidad et al. 2004 reported a small segregation of ∼3 km s⁻¹ in the radial velocity, with the north-eastern group mainly redshifted and the south-western one mainly blueshifted with respect to the ambient cloud velocity), we think that this is consistent with the masers to be located in the leading edge of projected bow shocks (Raga et al. 1997; Trinidad et al. 2013; see Appendix A).

Figure 4.

Positions, radial velocities, and proper motions of the H₂O masers associated with VLA 2 as observed with our set of three epochs of VLBA data (2001 December 2, 2002 February 11, and 2002 March 5). The colour scale indicates the radial velocity of the masers. The arrows represent the proper motion vectors of the H₂O maser features listed in Table 1. The solid arrows represent proper motions obtained from three epochs, while dashed arrows represent proper motions obtained from only two epochs. The (0,0) has absolute coordinates α(J2000) = 21^h43^m06|$.\!\!^{\rm s}$|323, δ(J2000) = 66°06′55|${^{\prime\prime}_{.}}$|92 (± 0|${^{\prime\prime}_{.}}$|05).

Open in new tabDownload slide

In order to extend the study of the evolution of the H₂O maser bipolar outflow in VLA 2 up to a time-span of ∼3.6 yr, we have aligned the H₂O maser positions obtained with the VLBA by Marvel (2005) on the single-epoch 2003 February 23, as well as those obtained with the VLA by Trinidad et al. (2004) on the single-epoch 1999 June 29, within the same reference frame of our aligned set of three VLBA epochs (2001 December 2, 2002 February 11, 2002 March 5) described in Section 2. For this alignment, we have first identified an H₂O maser feature clearly persisting in these five epochs. We identify the VLA feature number 4 listed by Trinidad et al. (2004) in their table 2 (V_LSR = −17.8 km s⁻¹), with the two close VLBA features numbers 3 and 4 listed in our Table 1 (V_LSR = −17.5 and −18.2 km s⁻¹, respectively), and the two close VLBA features 25 and 26 listed by Marvel (2005) in his table 4 (V_LSR = −17.8 and −17.7 km s⁻¹, respectively). Then, we corrected the mean position of the two features 25 and 26 listed by Marvel (2005), assigning it to the location expected for epoch 2003 February 23 in our reference frame, obtained from the extrapolation of the mean position and proper motions of the features 3 and 4 of our data for epoch 2001 December 2 (Table 1), assuming that they have moved with constant velocity during this time-span. This inferred spatial correction was applied to all the H₂O maser features of epoch 2003 February 23 detected by Marvel (2005). We also applied a similar correction method to the position of the VLA feature number 4 (and therefore to all the data of epoch 1999 June 29; Trinidad et al. 2004) assigning the position expected from the extrapolation of our VLBA data. This must be taken as a rough alignment, mainly because the VLA observations are made with a beam size of 80 mas, therefore, large enough to contain several of the individual features observed with the VLBA, together with the assumption of constant velocity of the maser features during the long time-span of ∼3.6 yr. In Fig. 5, we show the positions of the H₂O masers associated with VLA 2 measured with the VLA (epoch 1999 June 29) and VLBA (epochs 2002 February 11 and 2003 February 23). From this figure, we see that, overall, the whole structure of the H₂O masers as observed in epoch 1999 June 29 is contained within the structure observed in 2002 February 11, which in its turn, is contained within that observed in 2003 February 23, indicating that the structure has been continuously expanding during ∼3.6 yr. Measuring the angular separation between the north-eastern and south-western edges in 1999 June 29 (∼373 mas) and in 2003 February 23 (∼405 mas), we obtain an average expanding proper motion from a common centre of ∼4.4 mas yr⁻¹ (∼19 km s⁻¹). This estimate is roughly consistent with the proper motions we have measured for the individual H₂O maser features in the shorter time-span of ∼0.25 yr (Table 1 and Fig. 4).

Figure 5.

Positions of the H₂O masers associated with VLA 2 as measured with the VLA epoch 1999 June 29 (Trinidad et al. 2004), VLBA epoch 2002 February 11 (this paper), and VLBA epoch 2003 February 23 (Marvel 2005). The (0,0) has absolute coordinates α(J2000) = 21^h43^m06|$.\!\!^{\rm s}$|323, δ(J2000) = 66°06′55|${^{\prime\prime}_{.}}$|92 (± 0|${^{\prime\prime}_{.}}$|05).

Open in new tabDownload slide

Finally, with respect to the proper motions of the H₂O maser features associated with VLA 3A, VLA 3B, or VLA 3 (A/B) (Table 1), we find that they do not show a preferential motion pattern in the sky from which we can extract any helpful physical information. Therefore, in the next section, we will focus only on discussing the results we have obtained in VLA 2.

3 DISCUSSION

Our VLBA observations show a very compact, bipolar H₂O maser outflow centred on VLA 2, with the major axis of the maser distribution at p.a. ∼ 247° (see Fig. 4). Furthermore, the chain of masers to the south-west of VLA 2 can be fitted by a parabola. Using the offset positions of the H₂O maser features 1–20 in Table 1 that are shaping the south-western maser structure of VLA 2, one obtains a parabola which represents the data with a standard deviation of ∼15 mas. However, if we exclude features 12 and 13, the deviation is reduced significantly to ∼4 mas. On the other hand, the length of the chain of masers is ∼200 mas, so the average distance between the masers and the parabolic fit represents only 2 per cent of its length. The position angle (p.a.) of the axis of symmetry of the parabola is ∼230°, similar to the p.a. of the major axis of the full maser distribution around VLA 2. Another interesting result is that an extrapolation of the symmetry axis of the parabola passes at only ∼5 mas from VLA 2. If one forces the axis of the parabola to point exactly to this source, one obtains the following values for the best-fitting parameters: coordinates of the focus, x₀ (RA offset) = −119 mas, y₀ (Dec. offset) =−100 mas (±3 mas); distance between the focus and the tip of the parabola h = 22 ± 1 mas; and p.a. of the symmetry axis =230 ± 3° (see also Appendix A).

VLA 2 (NW2) is the weakest radio continuum source (∼0.1 mJy) detected by Trinidad et al. (2004) at cm wavelengths in the region. It is not seen in the J (1.25 μm), H (1.635 μm), and K (2.2 μm) bands, but it is seen in the L′ (3.77 μm) and M′ (4.68 μm) bands, and it appears bright in the mid-infrared (11.8 and 17.65 μm) bands, as reported by Kato et al. (2011), who conclude that this is an embedded YSO. This source is fainter in the near- and mid-infrared than the nearby source NW1, coinciding with VLA 3A+3B (see Fig. 1). These facts led Kato et al. (2011) to conclude that VLA 2 (NW2) is more embedded than NW1, or alternatively it has a lower luminosity. From the spectral energy distribution of NW1, Kato et al. (2011) estimate a luminosity of ∼700 L_⊙ for this object. These authors propose that NW1 is a YSO with spectral type B6-7, although this must be taken with caution, as it has been inferred assuming that it is a single main-sequence star, while NW1 appears to be an embedded binary since it has two components seen in radio continuum (VLA 3A and 3B) that are separated by ∼0.3 arcsec (∼270 au; Trinidad et al. 2004). Kato et al. (2011) did not determine the spectral type of VLA 2 (NW2). Here, using the VLA centimetre results, we discuss on the mass of VLA 2 under different assumptions. This source has a 3.6 cm flux density S_3.6 cm ≃ 0.1 mJy (Trinidad et al. 2004). Assuming that this emission arises from an optically thin photoionized H ii region, at a distance of d = 0.9 kpc, and with an electron temperature of 10⁴ K, the required rate of ionizing photons is ∼7 × 10⁴² s^{− 1} (e.g. Rodríguez et al. 1980). This ionizing photon rate can be provided by a B4 ZAMS star (Thompson 1984) with a luminosity of ∼700 L_⊙, which would correspond to a main-sequence star of ∼6 M_⊙ (Hohle, Neuhäuser & Schutz 2010). These estimates are likely upper limits given that part of the ionizing photons could be provided from shocks produced in the interaction of the outflow with circumstellar material (Torrelles et al. 1985; Curiel et al. 1989). Therefore, if we assume that the observed centimetre emission of VLA 2 traces a thermal radio jet, then ionization is expected to be produced by shocks (Anglada 1995, 1996). From the correlation between the observed S_νd² and bolometric luminosity of radio-jet sources (see fig. 5 in Anglada 1995), we infer a bolometric luminosity of ∼50 L_⊙ for VLA 2, which can be provided by an embedded protostar of a few solar masses (e.g. Wuchterl & Tscharnuter 2003; White et al. 2007). High angular resolution observations with the recently improved sensitivity of the VLA can determine whether VLA 2 shows the elongation and characteristics expected for a thermal radio jet.

From the spatial extension of the bipolar H₂O maser outflow, ∼ 0.2 arcsec (∼180 au; each side of the bipolar outflow, measured from VLA 2) and the expanding velocity of ∼20 km s⁻¹ measured through the radial and proper motions of the masers, we estimate a kinematic age of ∼40 yr for the outflow. Given how unlikely it would be to observe this short-lived outflow if it happened only once in the lifetime of the YSO VLA 2, we propose that the observed outflow does not represent a single, steady process, but rather it is part of a repetitive pulsed jet event. In this sense, Trinidad et al. (2004) proposed that VLA 2 is the powering source of the blueshifted [S ii] jet observed in the region and extending to the south-west with p.a. ∼ 252° (Ray et al. 1990), because this p.a. is similar to that of the major axis of the maser distribution measured with the VLA (p.a. ∼ 247°; Fig. 1). Our VLBA H₂O maser observations give strong support to this interpretation because in addition to the spatial distribution of the masers, the bipolar motions are along a similar direction (p.a. ∼ 247°; Fig. 4) as the [S ii] jet. Moreover, Fuente et al. (2001) propose that the redshifted counterpart of the blueshifted [S ii] jet is the ¹²CO (J=1–0) outflow observed towards the north-east, extending up to ∼100 arcsec with a kinematic age of ≳8×10³ yr. It might be possible that this extended outflow with higher kinematic age could be produced through repetitive short-lived ejections from VLA 2, as the one we find at small scales through our VLBA H₂O maser observations.

As mentioned in Section 1, short-lived episodic ejection events have been found towards massive YSOs through multi-epoch VLBI H₂O maser observations. With our VLBA observations, we have extended that kind of study to intermediate-mass YSOs, providing an example of a short-lived episodic ejection event associated with the source VLA 2 (NW2) in the star-forming LkHα 234 region. This could indicate that the accretion process in this YSO is also a non-steady process, but undergoing episodic increases in the accretion rate from the expected associated circumstellar disc, as it has been observed in low-mass YSOs through FU Orionis events or episodic ejection in the jets associated with HH objects (e.g. Hartmann & Kenyon 1996; Schwartz & Greene 2003; Lee et al. 2007; Greene, Aspin & Reipurth 2008; Pech et al. 2010; Stamatellos, Whitworth & Hubber 2012). Actually, episodic accretion has been proposed to solve the so-called luminosity problem in the modelling of Class I sources (Kenyon et al. 1990; Osorio et al. 2003).

Finally, our data, showing the presence of a compact (∼360 au) well-collimated bipolar outflow in VLA 2 (NW2), predict the presence of an accretion disc at these scales. In this sense, the central masers which are distributed along a linear structure with a size of ∼10 mas, first reported by Marvel (2005) and being more evident in the VLBA epoch of 2003 February 23 (see Fig. 5), could trace the innermost interaction region of the jet with its disc as suggested by this author. The presence of the disc can be tested through (sub)mm dust continuum and thermal line observations. These observations, together with sensitive radio continuum observations, would allow us to characterize not only the nature of the VLA 2 (NW2) object, but also the nature of the cluster of YSOs found in this intermediate-mass star-forming region.

4 CONCLUSIONS

We report VLBI H₂O maser observations towards the intermediate-mass star-forming LkHα 234 region. We have detected H₂O maser emission towards four YSOs (VLA 1, VLA 2, VLA 3A, and VLA 3B) within a region of ∼4 arcsec (∼3600 au) in size, indicating outflow activity from these objects. In particular, the spatio-kinematical distribution of the masers reveal a remarkable, very young, compact bipolar H₂O maser outflow associated with the intermediate-mass YSO VLA 2, which is located at the centre of the outflow. This outflow is formed by two bow-shock-like structures moving in opposite directions. The size of the outflow (0.2 arcsec = 180 au; measured from VLA 2), together with the expanding velocity of the masers with respect to the central source (∼20 km s⁻¹), gives a kinematic age of ∼40 yr. We interpret this outflow as driven by a short-lived episodic jet ejection from VLA 2, rather than a steady outflow. Short-lived episodic ejections have been also observed in massive YSOs through VLBI H₂O maser observations. With the results presented in this paper we show now that this behaviour may also occur in intermediate-mass YSOs, probably due to episodic increases in the accretion rate as observed and firmly established in low-mass YSOs. In this way, our observations predict the presence of an accretion disc around VLA 2 forming a disc–YSO–jet system similar to what is observed in low- and high-mass YSOs. This might be studied with sensitive radio continuum and (sub)mm dust continuum and spectral line observations at scales of a few tenths of an arcsec.

Our results support VLA 2 as the powering source of the extended outflow (∼100 arcsec = 9000 au) seen in [S ii]/CO. It is possible that this extended outflow, with kinematic age ≳8×10³ yr, might be driven by short-lived episodic ejections from VLA 2 as the very recent one traced by the VLBI H₂O maser observations.

We are very grateful to Eri Kato for providing us the near- and mid-infrared images shown in this paper (Fig. 1). We thank the referee, Kevin Marvel, for his valuable comments and suggestions on the manuscript. GA, RE, JFG, JMG, and JMT acknowledge the support from MICINN (Spain) AYA2011-30228-C03 grant (co-funded with FEDER funds). JC acknowledges support from CONACyT grant 61547. SC acknowledges the support of DGAPA, UNAM, CONACyT (México), and CSIC (Spain). CC-G and LFR acknowledge the support of DGAPA, UNAM, and CONACyT (México). MAT acknowledges the support from CONACyT grant 82543. JMG, RE, and JMT acknowledge the support from AGAUR (Catalonia) 2009SGR1172 grant. The ICC (UB) is a CSIC-Associated Unit through the ICE (CSIC).

REFERENCES

APPENDIX A: BOW-SHOCK APPROACH

Figure A1.

The dots show the proper motions |$v_{x^{\prime }}$|⁠, |$v_{y^{\prime }}$|⁠, and v_t of the individual H₂O maser features as a function of the y′ coordinate (measured perpendicular to the symmetry axis of the projected bow shock). The solid lines correspond to the least-squares fit described in the text.

Open in new tabDownload slide

The dispersion of the values of the fitting parameters that we obtain indicates that a ‘simple’ bow-shock model with a paraboloidal shape, although it can be considered as a rough representation of the observations, cannot reproduce exactly all the observed spatio-kinematical properties of the H₂O masers in the region.